Re: Please pull 'for-2.6.26' branch of 4xx tree

[Josh Boyer <jwboyer@linux.vnet.ibm.com>]

On Wed, 16 Apr 2008 22:50:47 +1000
Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote:

> 
> On Wed, 2008-04-16 at 07:31 -0500, Josh Boyer wrote:
> > On Wed, 16 Apr 2008 07:27:58 -0500
> > Josh Boyer <jwboyer@linux.vnet.ibm.com> wrote:
> > 
> > > On Tue, 15 Apr 2008 10:27:04 -0500
> > > Josh Boyer <jwboyer@linux.vnet.ibm.com> wrote:
> > > 
> > > > Valentine Barshak (2):
> > > >       [POWERPC] ibm_newemac: PowerPC 440GX EMAC PHY clock workaround
> > > >       [POWERPC] ibm_newemac: PowerPC 440EP/440GR EMAC PHY clock workaround
> > > 
> > > These two commits have been dropped from the branch.  They cause the
> > > allyesconfig build for powerpc to fail because CONFIG_PPC_DCR_MMIO
> > > doesn't understand the dcri_clrset macro.
> > 
> > Crap.  Looks like you already pulled the tree.  We need to either
> > revert those two commits, or essentially define dcri_clrset to a do {}
> > while(0) in include/asm-powerpc/dcr-mmio.h so things build.
> 
> Why ? It should do the right thing, not nop ...

"... so things build."  Relax Ben.  That isn't a final solution.

I'd rather just do an outright revert of the commits for now.

For a final solution, we're going to have to come up with a more clever
way of doing the clocking stuff in ibm_newemac.  The SDRs are board
specific and completely outside of EMAC/MAL itself.  So doing a dcr_map
on it becomes a bit more difficult, since there really isn't an SDR
node in the device tree at the moment.

josh